What I realized in the midst of the Kumamoto earthquake: what is missing from 'disaster prevention information'

First, to those in Kumamoto, and to those who have family or friends in Kumamoto.
I hear that aftershocks are continuing even as I write this article.
I imagine many of you are spending an anxious night, have been frightened by the shaking, or are feeling unsettled because you do not know the status of lifelines or transportation.
I offer my heartfelt sympathies.
Please, make your own safety and that of your family your top priority.

This article is a record of how I actually worked with AI in the midst of that situation.

To conclude, I did not make it in time for this earthquake.
Even so, I took up my pen because I believe there is value in recording my failures regarding two major challenges that I saw precisely because I took action:
“people who fall through the cracks of multilingual support” and
“the act of searching for information itself drains your battery.”

Chapter 1: On the day I thought, “Is there anything I can do?”

For about an hour after the earthquake occurred, I thought,
“Public institutions and companies take time to get approvals. But as an individual, couldn't I move much faster, even within the scope of personal responsibility?”

Since the Great East Japan Earthquake, various disaster prevention measures have progressed, but I wondered if there was anything I could do based on my own experiences as an individual of things I found difficult or wished were different.

Two specific ideas came to mind.

• Immediate fact-checking of rumors and misinformation

• Multilingual information conversion for foreign residents and tourists

While consulting with AI,
I realized that the mechanism in (1) where “AI determines truth or falsehood” carried too high a risk of misjudgment in life-or-death situations, so I changed my approach.
Instead of “determining,” I decided to “list links to primary sources (Japan Meteorological Agency, local governments, NHK, etc.) in a way that can be found quickly and without hesitation.”
—I judged that this was something I could shape immediately, even on my own responsibility.

Chapter 2: The public institution barrier of “15 languages”

The disaster prevention app “Safety tips,” supervised by the Japan Meteorological Agency and the Japan Tourism Agency, already supports 15 languages.
Japanese, English, Chinese (Simplified/Traditional), Korean, Spanish, Portuguese, Vietnamese, Thai, Indonesian, Tagalog, Nepali, Khmer, Burmese, and Mongolian.
Countries with many technical intern trainees are actually well-covered within this framework.
The government’s multilingual support is by no means negligent.

However, as I proceeded with the work, I noticed the existence of people who were quietly falling through the cracks of this framework.
Indian people running curry restaurants.
Ukrainian people and Russian speakers who have evacuated to Japan.
And... people like German tourists who happen to be in Japan for sightseeing and get caught in a disaster.

During the Great East Japan Earthquake, there was a story about a German parent and child who were visiting for sightseeing; they were caught in the tsunami in a taxi they had boarded to head to Sendai immediately after the earthquake, and although they managed to reach an evacuation center, they were at a loss with no one to rely on in a distant, foreign land.
(Afterward, it seems they were able to return home safely thanks to a relay of goodwill from people. *From Tohoku Gakuin University, “Record of the First Year of the Great East Japan Earthquake at Tohoku Gakuin: The Johanna Spielburg Family”.)
Tohoku Gakuin University “Record of the First Year of the Great East Japan Earthquake at Tohoku Gakuin”).
If people like that parent and child were in Kumamoto now, where should they look to find the correct information?
And I thought, if I were abroad and encountered such a disaster, and I had to pick up my smartphone, look for translation functions... how would I manage with a budget SIM or unstable communication?

It is natural for the government to operate within a “fixed framework” of 15 languages due to the system. That is precisely why, as an individual, I can flexibly add Hindi, Russian, Ukrainian, and German according to the needs on the ground, such as “there are many people from this country this time.”
I think this is not so much a difference in approval speed, but rather that the roles we can play are fundamentally different.

Chapter 3: The often-overlooked constraints of battery and communication

There is one more thing I keenly felt while working.
Many existing disaster prevention sites specialize in “displaying the current situation,” which is a correct design in itself.
However, from the perspective of the person affected by the disaster, while they are searching here and there to check the situation, their battery drains and communication becomes unstable.

During the Great East Japan Earthquake, I had the experience of “realizing later that charging a smartphone in a car was an act of consuming gasoline—in other words, a lifeline.”
The act of searching for information itself consumes limited and precious resources.

That is precisely why,
“If you just look here, you can reach various official information in one go”
—if such a lightweight and concise entry point had been prepared in advance and distributed in various places... This was what I felt most strongly while working on this.

Chapter 4: In the end, I didn't make it in time

The multilingual link collection itself took shape in a few hours.
However, there was a waiting time for the publication settings and activation on GitHub Pages, so I did not make it in time for the initial response to this earthquake.

Even if you think you are prepared, you may not be in time when the moment comes.
Honestly, this is a very frustrating experience.

However, if you don't take action,the frustration of 'not being in time'won't even arise.
Because I had a vision and actually put it into motion, I was able to see concrete issues that can be applied next time, such as:
'preparations for hosting'
'preventing cost spikes'
'a design philosophy of focusing only on primary information without making judgments'

Here is the page I actually published.

Japan is a country prone to earthquakes.
Even if I didn't make it in time for the initial response this time, I definitely want to make it in time for future responses to this disaster and for the next disaster.

If possible, it would be ideal if we could create a flow where volunteers in each region could create their own 'lightweight link collections that connect immediately to the latest information' based on this mechanism.
If information is not concentrated in one place but exists in a distributed manner on servers in various regions, it will be easier to view even when communication networks are congested.

There is a limit to what one person can do.
Even so, I believe that if the number of people creating this mechanism in each region increases little by little, disaster prevention information for all of Japan will become stronger.

I want to do what I can as an individual—because I believe that is my way of repaying the support I received from Japan and abroad during 3.11.

Chapter 5: 'Operations disappear, judgments remain'

This is the very theme of the Kindle book I am currently writing.

The 'operations' of the tool called AI will blend into the background and disappear like air as time goes by.
Just as we once learned mouse operations with Solitaire, no one will eventually be conscious of how to type search commands or even how to write prompts today.

However, only the 'judgment' of 'what to prioritize, what to discard, and how to act'judgmentwill remain across the ages.

During 3.11, what people in the disaster-stricken areas where communication was cut off were forced to make was precisely this 'judgment'.

This time, I hope that this record of failure—what we, who have gained the powerful 'right arm' of AI, created, published, and failed to do in time on our own responsibility—will become 'material for judgment' for someone else when they take action next.

Aiming for a society where such 'guideboards to primary information' are prepared in advance in each region, and everyone can recall, 'Oh, there was a summary site there.'
First, I published it, including my own failures.

Does your town have such a 'guideboard to primary information'?
Or, is there anything you thought 'it would have been good to have this' in a situation like this? Please let me know in the comments.

👉 [Kumamoto Earthquake Official Information Link Collection]

(*https://chikako921.github.io/Kumamoto_disaster_links/ )
If you don't mind, I would be happy if you could introduce it to those in Kumamoto or those who have acquaintances there.
It supports 12 languages, not just Japanese.
I have also linked information related to earthquakes, information from government offices, and pages related to JR and airlines.
For those whose languages are not supported by official sites, I have written that on the first page and recommend using English sites.